•All participants were monolingual and were from families classified as
either middle or low socioeconomic status (SES).
•The appropriate language version of the Expressive One Word Picture
Vocabulary Test (EOWPVT; Gardner & Gardner, 2000 or EOWPVT-SBE; Brownell,
2000 ) was administered.
Adults
•All AA participants and many CR adults obtained at least some
experience in college and were from families classified as either middle
or low SES.
•The frequently missed items from each test were presented to adults
within similar cultures for naming.
•Vocabulary knowledge increased as age increased.
•A high percentage ( > 50%) of all 3 groups mislabeled flying, lights, mail,
USA, time, and rodents.
–The children’s responses were confirmed as being culturally
acceptable responses.
Assessing Cultural and Linguistic Biases
in an Expressive Vocabulary Test
Samantha Wyatt, Kyna Fasnacht, Ruth Huntley Bahr, and Tempii Champion Communication Sciences and Disorders
ASHA 2006
“Building Bridges Through
Communication”
November 16-18, 2006
Miami Beach, FL
Cultural Knowledge
•Cultural knowledge encompasses the common knowledge, values,
and beliefs that are shared by a population or community.
•As children acquire more knowledge about the culture they live in,
they also acquire common vocabulary that is universally known.
•Social class can influence the development of cultural knowledge.
(Bradford and Harris, 2003).
–Children in lower SES families receive varying amounts of
linguistic input (i.e., informative resources) compared to other
children.
Vocabulary Development
*Two separate groups of Costa Rican children were tested.
•The assessments were administered according to the
published instructions outlined in the test manuals.
•The two test editions contained identical test items. For the
Spanish version, the examiner is instructed to skip
vocabulary items that are not appropriate for the Spanish
culture. The EOWPVT-SBE was standardized on school-
aged Mexican American students.
•Incorrect responses were separated for further qualitative
analyses.
•To analyze test items for cultural and/or linguistic bias,
Procedures
Analysis
Results: Distribution of Scores for AA Children
•Mean standard score = 87.54 (SD = 10.37) = Low normal
• When compared to the normative test sample, these scores
were significantly lower — one sample t(25) = – 6.130, p < .001.
• Five children scored at or below 78 (1.5 standard deviations
below the mean).
0
4
8
12
60-70 71-80 81-90 91-100 101-110
Range of Standard Scores
n
Results: Distribution of Scores for CR Children
•There is a negative skew to the data.
•No 9 year olds passed (mean = 61.2; SD = 7.6)
•9/19 eleven year olds passed (mean = 84.47, SD = 12.8)
• When compared to the normative test sample, these scores were
significantly lower — one sample t(28) = – 8.005, p < .001.
0
4
8
12
50-60 60-70 71-80 81-90 91-100 101-110
Range of Standard Scores
Comparison of Error Types for the AA Children
•Since subordinate responses predominated, these participants
appeared to possess adequate semantic knowledge of each word.
–The mental representations of this group of children do not appear
to be impoverished or have disordered expressive links (Hoover &
Storkel, 2005).
•The development of cultural knowledge ultimately reinforces the depth
of associations among vocabulary words (Bradford & Harris, 2003).
Presence of a Vocabulary Deficit?
•Participant performances suggest that there is cultural bias evident on
both versions of the EOWPVT. The result is an underestimation of
vocabulary knowledge.
–The Florida sample received a mean score that was lower than the
standardized sample, indicating the possibility of bias.
–The CR sample also performed poorly on the bilingual measure.
•Consideration of cultural knowledge is essential in the assessment of
vocabulary.
Variability in Naming Responses by CR Adults
33%
56%
11%
Tiempo (time)
Relojes (watches)
Desired Response = Time
Medidores de tiempo (measurers of time)
Comparison of Frequent Errors by the CR Group
Word 9 yrs. 11 yrs. Adults
Drinks 70% 68% 14%
Flying 70% 74% 77%
Lights 80% 63% 86%
Mail 100% 100% 77%
USA (map) 100% 84% 77%
Time 100% 63% 57%
Emotions 100% 26% 14%
Rodents 100% 79% 86%
Phonological Lexical Part-Whole Flops/Flip Flops
Group NMean Age (yrs.;mos.) SD (in yrs.)
AA 10 18;8 1.03
CR 739;3 14.48